An Approach for Domain Reduction with Data Dependence in Mutation Testing

摘要

As a testing strategy to evaluate the completenessof test cases, mutation testing has been identified as a "faultoriented"technique for unit testing, which is mainly used togenerate complete test cases. The path-oriented technique oftest data generation is a highly efficient technique whichimplements test data generation by building and solvingconstraint systems. Most of path-oriented generationtechniques only take control dependence among statementsinto consideration, which is to build constraint system byanalyzing control flow graph. However, it neglects theinfluence of data dependence among statements onconstraint system. Therefore, this paper improved test datageneration technique of domain reduction and proposed anew domain reduction method with data dependence. Itadded detecting of equivalent mutants and solved influenceson constraint systems caused by multiple conditional branchstatement. Experimental results showed that this methodimproved success rate and execution efficiency of test datageneration in a significant extent.